System and method for dynamically expanding virtual cluster and recording medium on which program for executing the method is recorded

• US 9,571,561 B2
• Filed: 12/27/2013
• Issued: 02/14/2017
• Est. Priority Date: 12/28/2012
• Status: Active Grant

• Current Assignee: Samsung SDS (China) Co., Ltd. (Samsung Group)
• Original Assignee: Samsung SDS (China) Co., Ltd. (Samsung Group)
• Inventors: Jang, Sung Ho
• Primary Examiner(s): Chea, Philip
• Assistant Examiner(s): Chen, Wuji
• Application Number:
  US14/141,784
• Publication Number:
  US 20140189109A1
• Time in Patent Office: 1,145 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 2209/5022 Workload threshold
• CPC Class Codes: G06F 9/5077 Logical partitioning of resources; Management or configuration of virtualized resources specific details on emulation or internal functioning of virtual machines G06F9/455
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46

Abstract

Provided are a system and method for dynamically expanding a virtual cluster having one or more virtual machines (VMs), based on the resource availability of the virtual cluster and the type of the virtual cluster. The system for dynamically expanding a virtual cluster having one or more virtual machines (VMs), the system comprising a monitor unit which measures resource availability of a target virtual cluster and provides the resource availability to an expansion control unit, the expansion control unit which determines whether to expand the target virtual cluster based on the resource availability and determines a type and number of VM instances to be additionally allocated to the target virtual cluster by reflecting the resource availability and a type of the target virtual cluster and a virtual cluster configuration unit which modifies profile information of the target virtual cluster such that the determined number of VM instances of the determined type by the expansion control unit can be additionally allocated to the target virtual cluster.

Claims

1. A system for dynamically expanding a virtual cluster having one or more virtual machines (VMs), the system comprising:

a monitor measures resource availability of a target virtual cluster; and

a hardware processor configured to execute;

an expansion controller which receives the measured resource availability, determines whether to expand the target virtual cluster based on the measured resource availability and determines a type and a number of VM instances to be additionally allocated to the target virtual cluster based on the measured resource availability and a type of the target virtual cluster; and

a virtual cluster configurer which modifies profile information of the target virtual cluster such that the determined number of VM instances of the determined type can be additionally allocated to the target virtual cluster,

wherein the type of the target virtual cluster indicates which hardware resource is reinforced by the target virtual cluster, and

wherein the hardware resource is reinforced indicates performance of which hardware resource is better than performance of standard hardware resources, and

wherein the monitor measures utilization rates of first through n-th resources in the target virtual cluster and measures the resource availability using the utilization rates of the first through n-th resources, wherein, U_AVG=a1×

(1−

U_1)+a2×

(1−

U_2)+ . . . +an×

(1−

U_n), where U_AVG indicates the resource availability, U_1 through U_n respectively indicate the utilization rates of the first through n-th resources, a1+a2+ . . . +an=1, and a1, a2, a3, . . . , an are weights for the first through n-th resources, respectively.

2. The system of claim 1, wherein the expansion controller comprises:

a threshold adjustment module which automatically adjusts a threshold for the target virtual cluster based on at least one from among a number of VM instances in the target virtual cluster, a ratio of the sum of access times to an active period of the target virtual cluster, and a ratio of the number of VM instances allocated to a total number of VM instances that can be provided by a server pool in a virtual cluster service system; and

an expansion determination module which compares the measured resource availability with the threshold and determines to expand the target virtual cluster when the measured resource availability is less than the threshold.

3. The system of claim 2, wherein the monitor further measures the utilization rates of a first resource and a second resource of the target virtual cluster and provides the measured utilization rates to the expansion controller, and the expansion controller further comprises a required resource amount calculation module which calculates an additionally required amount of the first resource and an additionally required amount of the second resource when the expansion determination module determines to expand the target virtual cluster,

wherein ADD_1=(TH−

U_AVG)×

(1−

U_1)×

(1+a), where ADD_1 indicates the additionally required amount of the first resource, TH indicates the threshold, and a is a preset weight exceeding zero when the type of the target virtual cluster is a first resource-oriented type and is zero when the type of the target virtual cluster is not the first resource-oriented type, and ADD_2=(TH−

U_AVG)×

(1−

U_2)×

(1+b), where ADD_2 indicates the additionally required amount of the second resource, TH indicates the threshold, and b is a preset weight exceeding zero when the type of the target virtual cluster is a second resource-oriented type and is zero when the type of the target virtual cluster is not the second resource-oriented type.

4. The system of claim 3, wherein the expansion controller further comprises a VM selection module which determines the type and the number of VM instances that are to be additionally allocated such that the additionally required amount of the first resource and the additionally required amount of the second resource calculated by the required resource amount calculation module can be met,

wherein the VM selection module determines the type and the number of VM instances that are to be additionally allocated such that a minimum number of VM instances can be additionally allocated and that a maximum number of VM instances of a type corresponding to the type of the target virtual cluster can be additionally allocated.

15. The system of claim 1, wherein the type the target virtual cluster is selected from among a CPU-oriented type, a memory-oriented type, and a storage oriented type.

16. The system of claim 1, wherein the expansion controller determines the type and the number of the VM instances to be additionally allocated to the target virtual cluster, which is an existing cluster, by assigning different weights to different types of resources based on the type of the target virtual cluster that comprises a plurality of hardware resources.

5. A method of dynamically expanding a virtual cluster having one or more VMs, the method comprising:

determining resource availability of a target virtual cluster by using a virtual cluster service system;

determining whether to expand the target virtual cluster based on the determined resource availability by using the virtual cluster service system; and

additionally allocating one or more VM instances to the target virtual cluster by using the virtual cluster service system when it is determined to expand the target virtual cluster,

wherein the additionally allocating of the VM instances comprises;

determining a type and a number of VM instances that are to be additionally allocated, based on the determined resource availability and a type of the target virtual cluster; and

additionally allocating the determined number of VM instances of the determined type to the target virtual cluster,

wherein the type of the target virtual cluster indicates which hardware resource is reinforced by the virtual cluster,

wherein the hardware resource is reinforced indicates performance of which hardware resource is better than performance of standard hardware resources, and

wherein the determining of the resource availability comprises;

measuring utilization rates of first through n-th resources in the target virtual cluster; and

measuring the resource availability,

wherein U_AVG=a1×

(1−

U_1)+a2×

(1−

U_2)+ . . . +an×

(1−

U_n), where U_AVG indicates the resource availability, U_1 through U_n respectively indicate the utilization rates of the first through n-th resources, and a1+a2+ . . . +an=1.

6. The method of claim 5, wherein a1, a2, a3, . . . , an are a weight set corresponding to the type of the target virtual cluster.

7. The method of claim 5, wherein the determining of whether to expand the target virtual cluster comprises:

automatically adjusting a threshold for the target virtual cluster using the virtual cluster service system based on at least one from among a number of VM instances in the target virtual cluster, a ratio of the sum of access times to an active period of the target virtual cluster, and a ratio of the number of VM instances allocated to a total number of VM instances that can be provided by a server pool in the virtual cluster service system; and

determining to expand the target virtual cluster when the determined resource availability is less than the automatically adjusted threshold.

8. The method of claim 7, wherein the automatically adjusting of the threshold comprises increasing the threshold based on a first weight in a preset base threshold when automatically adjusting the threshold based on the number of VM instances in the target virtual cluster, wherein the first weight is a value that increases as (VM_MAX−

VM_USER)/(VM_MAX−

VM_MIN) increases, where VM_MAX is a number of VMs in a virtual cluster having a largest number of VMs among virtual clusters serviced by the virtual cluster service system, VM_MIN is a number of VMs in a virtual cluster having a smallest number of VMs among the virtual clusters serviced by the virtual cluster service system, and VM_USER is a number of VMs in the target virtual cluster.

9. The method of claim 7, wherein the automatically adjusting of the threshold comprises increasing the threshold based on a second weight in the preset base threshold when automatically adjusting the threshold based on the ratio of the sum of the access times to the active period of the target virtual cluster, wherein the second weight is a value that increases as the sum of the access times divided by the active period of the target virtual cluster increases, where the sum of the access times is a sum of session connection times of a user account of the target virtual cluster to the virtual cluster service system, and the active period of the target virtual cluster is a period of time from a generation of the target virtual cluster to a present time.

10. The method of claim 7, wherein the automatically adjusting of the threshold comprises reducing the threshold by reflecting a third weight in the preset base threshold when automatically adjusting the threshold by reflecting the ratio of the number of VM instances allocated to the total number of VM instances that can be provided by the server pool in the virtual cluster service system.

11. The method of claim 5, wherein the determining of the resource availability comprises:

measuring the utilization rate of the first resource and the utilization rate of the second resource in the target virtual cluster and calculating the resource availability of the target virtual cluster using the utilization rate of the first resource and the utilization rate of the second resource, and

the determining of the type and the number of VM instances that are to be additionally allocated comprises;

calculating an additionally required amount of the first resource using the resource availability and the utilization of the first resource and adjusting the additional required amount of the first resource based on a weight for the first resource according to the type of the target virtual cluster;

calculating an additionally required amount of the second resource using the resource availability and the utilization rate of the second resource and adjusting the additionally required amount of the second resource based on a weight for the second resource according to the type of the target virtual cluster; and

determining the type and the number of VM instances that are to be additionally allocated such that the additionally required amount of the first resource and the additionally required amount of the second resource can be satisfied.

12. The method of claim 5, wherein the determining of the type and the number of VM instances that are to be additionally allocated comprises:

calculating an additionally required amount of a resource using the determined resource availability; and

determining the type and the number of VM instances that are to be additionally allocated such that the additionally required amount of the resource can be met, wherein the type and the number of VM instances that are to be additionally allocated are determined such that a minimum number of VM instances can be additionally allocated.

13. The method of claim 5, wherein the determining of the type and the number of the VM instances that are to be additionally allocated comprises:

calculating an additionally required amount of a resource using the determined resource availability; and

determining the type of VM instances that are to be additionally allocated to be a type corresponding to the type of the target virtual cluster; and

determining the number of VM instances that are to be additionally allocated such that the additionally required amount of the resource can be met.

14. The method of claim 5, wherein the determining of whether to expand the target virtual cluster comprises:

periodically determining resource availabilities;

accumulating the periodically determined resource availabilities;

analyzing a trend of the accumulated resource availabilities;

predicting whether the determined resource availability will be less than the threshold after a predetermined number of periods based on the trend of the accumulated resource availabilities; and

determining to expand the target virtual cluster when it is predicted that the determined resource availability will be less than the threshold after the predetermined number of periods.